Preparation and characterization of activated montmorillonite clay supported 11-molybdo-vanado-phosphoric acid for cyclohexene oxidation

A new heterogeneous catalyst (PVMo/bentonite), consisting of vanadium substituted polyphosphomolybdate with Keggin structure H4[PVMo11O40]·13H2O (PVMo) supported on acid activated bentonite (clay from Hammam Boughrara, Maghnia, Algeria), was prepared by witness impregnation and characterized by X-ray diffraction, BET, Fourier-transformed infrared spectroscopy, 31P NMR, UV-vis diffuse reflectance spectroscopy and thermogravimetric & differential thermal analysis (TG-DTA). X-ray diffraction indicated that PVMo was properly loaded on bentonite as a support. Heterogenization of homogenous catalysts is really interesting, as heterogeneous catalysts are recoverable. Therefore, the synthesized materials can be used as efficient heterogeneous catalysts for epoxidation of cyclohexene. The obtained results showed that a better catalytic activity can be obtained with PVMo/bentonite (81.5% of conversion), by drop addition of H2O2, for 3 hours.

-ponding to POMs, namely heteropolyacids(HPAs), are known to possess a Brønsted acidity which is stronger than that of many mineral acids or conventional acidic solids; indeed, they can display a strong oxidative power.For these reasons, these clusters are widely investigated, to see whether they can be used as catalysts in reactions requiring both acidic and oxidative conditions; these reactions can be performed in homogeneous as well as heterogeneous systems [12][13][14][15][16] .
Vanadium incorporated molybdophosphoric acid (PVMo) catalysts show unique catalytic features for oxidation due to their bifunctional character which arises from the redox nature of vanadium and the oxidative/acidic character of molybdophosphoric acid 17 .Several authors have investigated V incorporated molybdophosphoric acid catalysts by replacing 1-3 Mo atoms by the corresponding number of V atoms 18,19 .The acidic, redox and thermal stability properties of PVMo catalysts are influenced by the degree of V substitution 20 .However, owing to their small surface area (< 10 m 2 g -1 ) and separation problems from the reaction mixture, their catalytic performances are often limited to heterogeneous catalysis 21 .For that reason, several authors proposed to disperse them on porous supports with large surface areas to make the active sites of HPAs more accessible to reactants.The deposition of the modified HPAs on supports is important to increase the surface area.It is always advantageous to use supports like silica, ZrO 2 , active carbon, SBA-15, and zeolite, etc... [22][23][24][25][26][27] .
The synergism of clay minerals and heteropolyacids was discussed for the development of many green processes with Page 1 of 13 RSC Advances

RSC Advances Accepted Manuscript
This journal is © The Royal Society of Chemistry 2014 potential industrial applications 28 .Acid modified clay minerals can also be used as efficient supports 29 because they exhibit higher surface area, pore volume, pore diameter and higher surface acidity 30 ; this gives them better adsorption and catalytic properties.In particular, acid treated clay catalysts have received considerable attention in different organic syntheses, as catalysts or as supports, because of their environmental compatibility, low cost and simplicity of use.Using clay catalysts, we can conduct environmentally benign green chemistry both at industrial level and at a laboratory scale.The acid strength of HPA, supported on some of the materials mentioned above, is lower than that of bulk HPA; this is due to the interaction of HPA with surface functional groups of supports.In the case of a composite material, the support has an influence on the acidity function of HPA and vice versa 31 .Acid-treated Montmorillonite exhibits relatively strong acid sites 32 .Several research groups have reported the use of activated clay-supported HPA (e.g., 12tungstophosphoric acid) in the etherification of phenethyl alcohol with alkanols 33 , synthesis of 1,5-benzodiazepine derivatives both in solution and under solvent-free conditions 34 , etherification of acetic acid with sec-butanol 29,35 , and hydroxyl alkylation of phenol, respectively.Beside HPA, many metals or metal oxides can be strategically supported on activated clay, metal oxide pillared interlayer clays (PILCs), and porous clay heterostructures (PCHs) as catalysts for organic synthesis 36,37 .
In the present study, we report that HPA like 1-vanado-11molybdophosphoric acid (PVMo) is supported on acid-activated Montorillonite clay having high surface area, prepared from Montmorillonite by HCl acid activation; it is also considered as a solid acid catalyst for the epoxidation of cyclohexene, using hydrogen peroxide as an oxidant.There are various reasons for choosing this reaction as a model from a wide variety of acid catalyzed and oxidation reactions.Cyclohexene epoxidation is attracting both academic and industrial interest in homogeneous and heterogenous systems 5,8,9,[38][39][40][41][42][43][44][45][46][47][48][49] .Cyclohexene epoxide is a valuable organic intermediate, used in the synthesis of products such as chiral pharmaceuticals, pesticides, epoxy paints, rubber promoters, dyestuffs, plant-protection agents and stabilizers for chlorinated hydrocarbons 43,[47][48][49][50][51] .Again, HPA supported on clay is never used in this reaction.The active centers, responsible for the reactions, are discussed in detail in the manuscript.

Catalyst preparation
The synthesis of

Characterization of catalysts
The composition and weight percentages of P, Mo, and V in the samples were evaluated by the XRF technique using a Thermo Electron ARL-ADVANT¨XP X-ray fluorescence spectrophotometer.
Prior to analysis, the instrument was calibrated against standard samples using the calibration kit (CAL-SS18) provided by thermo Scientific.Quantization was carried out using Thermo Electron UNIQUANT software.
The surface area and porosity were measured at liquid nitrogen temperature, using Quantachrome Instruments (Nova 1000 e ).Prior to each adsorption-desorption measurement, the sample was degassed at T = 150 °C for 3 h.The specific surface areas were determined using the BET equation and the microspore volumes by the de Boer's t-method.The α s plots were used to estimate the microporosity where α s = (n/n s ) ref (n s ) is the amount adsorbed by the reference solid (p/p 0 = s).
The materials were characterized by X-ray powder diffraction (XRD) using a Rigaku D/max2500 diffractometer with Cu Kα radiation (λ = 1.541874Å) in the range 2θ = 2° -70° with a step of 0.02° and an acquisition time of 1 s.
Fourier transform infrared (FTIR) spectra were recorded by an Agilent Technologies Cary 600 series FTIR spectrometer using KBr pellets, under atmospheric conditions.The FTIR spectra were obtained in KBr pellets in the range of 400-4000 cm -1 , and all spectra were collected at room temperature with a resolution of 4 cm -1 . 31P MAS NMR (161.70 MHz) spectra were measured at room temperature on a Bruker Avance 400 spectrometer, where 85 % H 3 PO 4 was used as an external reference.
The UV-vis diffuse reflectance spectra of the samples were collected in a JASCO V-670 UV-Vis spectrophotometer equipped with a Harrick Praying mantis diffuse reflectance accessory.Spectral grade BaSO 4 was used as a reference material.
The thermogravimetric and differential thermal analyses (TG-DTA) measurements of representative samples were carried out using a TA instruments 2960 SDT V3 thermal analyzer.The TG-DTA experiments were performed under argon and nitrogen.About 10mg of the sample was heated at a rate of 10 °C min -1 .Quartz cells were used as sample holders with α-Al 2 O 3 as a reference.The studied temperature range was 25-800 °C.

Evaluation of catalytic activity
Catalytic epoxidation reactions were performed in a round bottom flask, in an oil bath, under magnetic stirring (800 rpm).Unless otherwise specified, all oxidation reactions were carried out at atmospheric pressure under reflux with 30 mmol of cyclohexene; 60 % H 2 O 2 (30 mmol) was used as oxidant and was slowly added (3 h) with a syringe pump, 10 mL of acetonitrile as solvent, 1 mL of 1,2 dimethoxyethane as internal standard for product quantification, and 25 mg of catalyst.The reaction mixture was stirred at 70 °C.The catalytic reactions were monitored using gas chromatography (Agilent Technologies 7890A GC System) equipped with a Phenomenex Zebron-HT-5 Inferno (30 m x 0.25mm 0.25 µm) column; a flame ionization detector (FID) was also used.The conversion, product selectivity and product yield were calculated as follows: 3 Results and discussion

Chemical analysis (XRF)
The results of chemical analysis of heteropolyacids (Table 1) were adjusted considering 11 atoms of molybdenum per Keggin unit, according to the nature of HPA, and were found to be in good agreement with the desired stoichiometry for phosphorous, molybdenum and vanadium.The highest proportions of SiO 2 and Al 2 O 3 agree with the nature of the clay support.Moreover, for the impregnated samples, the initial Si, P, Mo and V amounts in PVMo/HMont were found again (within 1 wt.% experimental errors) at the end.Referring to the Mo/P and V/P ratios, average values of 11 and 1, respectively, were obtained for all three preparations, indicating that the quantities of metals used were preserved in the three supported samples.

Surface area and porosity
The surface area and pore volume of HMont and PVMo loaded HMont were measured.It appeared that, in general, the surface area and pore volume of HMont decreased considerably upon loading with PVMo.The results of a typical example, i.e.PVMo/HMont catalysts, are given in Table 2.The surface area (195 m 2 g -1 ) and the total pore volume of HMont decreased as the amount of PVMo loading increased.Thus, for 20%PVMo loading, the surface area fell to 141 m 2 g -1 and the total pore volume to 0.17 cm 3 g -1 only.The high surface area of acid activated clay was due to the removal of aluminum from the octahedral sites in the clay sheet.
The reduction in surface area and pore volume upon PVMo loading might be due to the blockage of pores by PVMo molecules.These observations suggest that the Keggin units do not penetrate inside the pores but rather stay outside of the surface of the support.Although, in most investigations, the authors considered that HPA can enter into the pores of the support 14,54,55 , we believe that these conclusions do not take into account the question of anion-anion repulsion and the chemical interaction between the support and HPA protons that could affect the final location of HPA in or on the support.
The complete adsorption-desorption isotherms of calcined samples, are presented in Fig. 1(A).All the curves are nearly similar and can be included in type IV in the BDDT classification.The hysteresis loops, at low pressure, indicate the presence of microspores [56][57][58] .Assuming the pores to be cylindrical, the average pore radius can be calculated using the formula r =2 V P /S p , where r is the average pore radius, V p is the pore volume, and S p is the specific internal surface area of the pores.The average pore radius calculated by this method is reported in Table 2.This shows that with the increase in % loading of PVMo, the average pore radius increases.It is interesting to note that the pore size distribution, determined from the desorption branch of the N 2 adsorption isother using the Barrett-Joyner-Halenda (BJH) method (Fig. 1B), shows that the average pore sizes of different weight percents of PVMo on HMont are in the microporous region.The decrease in intensity of the diffraction peaks, attributed to clays, was already observed on HMont 29 , Mont K10 60 and silica pillared clay 61 in the presence of heteropoly acids.Moreover, no pattern of the PVMo crystal phase, after impregnation of HPA, indicates that HPA is finely dispersed on the HMont surface; PVMo particles are too small and/or too well dispersed and therefore undetectable by XRD.  62,63After the materials were supported on HMont, some of the characteristic Keggin bands were observed at 982 cm −1 and 893 cm −1 and others merged up with the HMont bands.
A decrease in the intensities of these bands was noted as the loading increased.

31 P NMR study
The solid state 31 P (MAS) NMR of PVMo is shown in Fig. 5.The 31 P chemical shift provides important information about the structure, composition and electronic states of these materials.The 31 P NMR spectra of PVMo (Fig. 5) show a chemical shift at -3.884 ppm; this is in good agreement with the reported one 23,59,64 .

UV-vis diffuse reflectance spectroscopy
DRUV-Vis spectroscopy provided additional information about the substitution of ions into the Keggin units.The charge transfer absorption spectra of most unreduced polyanions appear in the 200-500 nm range.The acid activated Montorillonite clay, without HPA, displays a characteristic broad band centred at about 245 nm (Fig. 6a) 65,66 .showed absorption maxima at 258 and 308 nm.The 258 nm absorption is typical for the Keggin structure of H 4 PMo 11 O 40 28 , while the 308 nm shoulder is characteristic of Vanadium incorporation into Keggin ion 1,51 .These bands are due to ligand-to-metal charge-transfer (LMCT) transitions associated with octahedrally coordinated Mo 6+ units 67 .These two absorption bands are clearly seen in the solid-state UV-vis spectra of 20%PVMo, while in the 10%PVMo, the 308 nm band is hardly seen.Since pure HMont (Fig. 6d) shows no UV absorption, then this result confirm the presence of the undegraded primary Keggin structure of PVMo species in 20%PVMo, but suggests a lack of Vanadium substituted HPA species in the case of 10%PVMo.

Differential thermal analysis and thermo gravimetric analysis (DTA/TGA)
The DTA/TGA curves of PVMo, HMont, 10%PVMo/HMont and 20%PVMo/HMont are presented in Fig. 7. Between 50 and 125 °C, DTA shows that all water molecules do not play the same role.In this temperature range, endothermic peaks due to the presence of water molecules interacting in different ways were observed (Fig. 7A).Some of them within clusters, such as H 3 O + or H 5 O 2 + , are more tightly bound in the crystal structure.HPA possesses the crystallization water that binds the close Keggin units in the secondary structure by forming water bridges.The weight loss, between 50 and 125 °C, observed in the TGA analyses, indicates that PVMo is hydrated with 13 molecules of water (Fig. 7A).Above ~ 450 °C, DTA shows that HPAs decompose in a mixture of oxides P 2 O 5 , V 2 O 5 and MoO 3 62, 68 .The DTA/TGA of acid activated clay HMont (Fig. 7B) showed an endothermic signal below 100 °C which represents the dehydration of the smectic structure.A steady loss of mass of about 17 %, up to 500 °C, was attributed to the loss of physisorbed and interlayer water and to the dehydroxylation caused by the breaking of structural OH-groups in the support 30,69 .The TGA of supported PVMo onto HMont (Figs. 7C and 7D) showed a mass loss of about 18 % within the temperature range of 80 -150 °C due to a loss of adsorbed water molecules.A gradual mass loss of about 3 %, up to 500 °C, was also observed, which indicates an increase in the thermal stability of PVMo on HMont support.This might be due to the formation of intermolecular bondings between the support and the heteropoly acid; it indicates the presence of a chemical interaction between them 35,70,71 .

Epoxidation reaction
Cyclohexene oxidation with H 2 O 2 60 % was chosen as a model reaction to compare the behaviours of different catalysts.
Cyclohexene epoxidation with alkyl-hydroperoxides mainly yields cyclohexene oxide (epoxide), however when the oxidant is hydrogen peroxide some other products are observed (Fig. 8).
Trans-1,2-cyclohexanediol is obtained by the acid catalyzed opening of the epoxide with water, small amounts of cyclohexanone (one) can be obtained by the dehydration of this diol with strong acids.Allylic oxidation is a radical process leading to cyclohexene-hydroperoxide, which acts as a secondary oxidant leading to the epoxide and cyclohexenol, furthermore this product dehydrates to yield cyclohexenone, in fact it dehydrates in the injector of the GC so that this amount in the mixture cannot be determined, so the amount on cyclohexenol (enol) and cyclohexenone (enone) determines the contribution of the allylic oxidation.In order to reduce both side-reactions and non-produtive decomposition of the oxidant, a large excess of alkene is used in most of cases, as we want to determine the influence of several parameters in the reaction results we have used equimolecular amounts of both reagents.In order to highlight the impact of heterogeneous catalysts, homogeneous catalytic tests were also achieved.The effects of the nature of H 2 O 2 , catalyst percentage and reaction time are taken into consideration.
The results show that the way H 2 O 2 60 % is added has an impact on conversion (Fig. 9).The best conversion (81.5 %) is reached when hydrogen peroxide is drop-wise added during 3 hours, selectivity to 1-2-diol ids very high (91.6 %) with a small contribution (4.6 %) of allylic oxidation.
The catalytic activities of different samples were evaluated for the oxidation of cyclohexene.All the catalytic systems proved to be highly active and selective towards diol.A blank reaction was performed in the absence of catalyst and no product was detected.
Similar oxidation reactions were also carried out with acid activated montmorillonite support and only a negligible catalytic activity was evidenced, and also no desired product was detected.
The obtained results clearly indicate that the POM species are responsible for the high oxidation activity of these heterogeneous catalysts.Besides, the large pores of the HMont support permit diffusion of a relatively large quantity of cyclohexene and oxidation proceeds selectively inside the pores of the support.Generally, the reaction is believed to proceed through the mechanism postulated for Keggin type POMs.Monomeric, dimeric, and tetrameric peroxo species are generated by the reaction of polyoxometalates with hydrogen peroxide, and the peroxo species are supposed to be the active species for epoxidation of olefins within a POM/H 2 O 2 system 42,67,72 .It has been postulated that the epoxidation reaction rate and epoxide yield, for the same type of catalytic species, are determined by the catalyst's acid strength 40,73 .The lower activity of the PVMo/HMont, in comparison with PVMo catalyst, shows the importance of diffusional limitations.However the high diol selectivity with 20% PVMo/HMont is clearly justified considering the acidity of the support.The catalytic performance of the heterogeneous catalysts is also related to the fine dispersion of the active catalytic species within the supports.
Here, the acid Brønsted sites of PVMo, acid Lewis sites of support and redox properties of PVMo are shown to have a crucial effect on the oxidation process.From table 3, it can also be seen that the catalytic activity of 20%PVMo-HMont is better than that of 10%PVMo-Hmont.The importance of the support has been also observed using HPAs immobilized in other supports, such as such as MCM-41 24 SiO 2 , SBA-15, ZrO 2grafted SBA-15 74 .The comparison of the reaction at different temperatures (Table 4) so that, with both supported catalysts, best yields and selectivity's are reached at 70 ºC, both results are worse in reactions carried out at lower and higher temperatures.

RSC Advances Accepted Manuscript
A study was also carried out in order to see the effect of reaction time on the catalytic activity, with 0.025 g of catalyst, 30 mmol of cyclohexene, 30 mmol of hydrogen peroxide, 10 mL of acetonitrile, at the temperature 70 °C.The experimental results (Fig. 10) indicate that the reaction proceeds more slowly in the presence of heterogeneous catalysts.This may be attributed to diffusional restrictions of the reagents inside the pores of the catalyst.The energy of activation was calculated using rate constant (k) determined at different temperatures.It is a well-established fact that in the oxidation of hydrocarbon, the rate of the reaction is proportional to the product concentration.The conversion vs time plots obtained in the present work indicate that the cyclohexene oxidation is not an exception to this fact.The reaction is considered to be second order with respect to the product and reactant concentration.The rate expression used by the many workers for oxidation reactions 2, 10, 11 is adopted here.

−d[cyclohexene]/d(t) = k[cyclohexene] 2
(1) on integrating expression (1), from the initial concentration at initial time to the final concentration at the final time t , the expression can be written as : (2) A plot of 1/[cyclohexene] with respect to time gives a linear relationship showing a pseudo-second-order dependence on the cyclohexene, k represent the apparent rate constant.Fig. 11a shows this plot.All the points fall on a straight line which proves that the rate expression holds good under the conditions employed for the reaction in the present work.Gupta et al and M. Abrantes et al 2,3 reported the same apparent activation energy.However this value was low to the values reported in many works (Table 5).

Catalyst reusability
As the reusability of a catalyst is important from the economical and industrial points of view, we decided to investigate the reusability and stability of PVMo/Hmont in oxidation reactions.The catalysts were filtered and reused after 9 h of reaction time.They were washed for several times with a solvent, and then dried at room temperature and finally used in cyclohexene oxidation with a fresh reaction mixture.As a typical example, the catalyst 20% PVMo / Hmont showed (Fig. 12) a conversion of 81% in the first run, which decreased to about 61, 48 and 39 % in the 2 nd , 3 rd and 4 th run, respectively.The amount of catalyst leaching after each run was determined by ICP analysis.In this manner, the filtrates were collected after each run and used to determine the amounts of leached vanadium (Table 6).The decrease in catalytic activity is due to the leaching of HPA (of about 1-2 %) from the support into the liquid phase during the catalytic reactions; this was confirmed by the well-known ascorbic acid test 75 .Although the catalyst was washed after filtration to remove all the adsorbed reactants and products, retention of some of the adsorbed reactants and products was still possible, and this might cause the blockage of active sites and consequently the decrease in the catalytic activity.The nature of the recovered catalyst was followed by FT-IR.After reusing the catalyst for several times, no change in the FT-IR spectrum was observed.However, the selectivity remained almost unchanged, and the catalysts were therefore reused four times.

Conclusion
In this work, we successfully prepared and characterized PVMo/HMont by a simple impregnation method.The XRF analysis showed that the procedure used gives the expected materials, and the PVMo structure is confirmed by 31 P NMR analysis.UV-Vis and IR bands show the Keggin structure of HPA.XRD analysis showed that PVMo is well dispersed on the surface of Montmorillonite.Thermal analysis showed that the material prepared is stable for temperatures up to 125 °C; this is explained by the formation of intermolecular bonds between PVMo and the support.The prepared heterogeneous catalysts showed high catalytic activity for cyclohexene epoxidation.They are not expensive and can be used in heterogeneous media.These catalysts showed that the reaction is oriented rather towards epoxides than enol.The use of a very small amount (25 mg) enabled to reach a good conversion (above 80 %) and yield of diol (about 77 %).The use of homogeneous catalyst PVMo gives only 64 % of diol.Also, using drop addition for 3 hours increased the conversion.H 2 O 2 is a green and eco-friendly oxidant in these systems.

Fig. 11 a
Fig. 11 a second-order kinetic plot for cyclohexene epoxidation with hydrogen peroxide over 20%PVMo-Hmont Fig. 11b displays the graph of ln (k) obtained from expression (2) within 6 h versus the inverse of the reaction temperature for each temperature.The apparent activation energy (Ea) calculated from the Arrhenius plot (Fig.11b) was ca.3.216 kcal mol -1 .K.C. Gupta et al and M. Abrantes et al2,3 reported the same apparent activation energy.However this value was low to the values reported in many works (Table5).

Table 1
59emical composition of catalysts from XRFThese numbers correspond to the numbers of atoms deduced from the raw wt.% found from XRF, as the number of Mo atoms is fixed to 11.The XRD patterns of supports HMont, PVMo, 10 and 20%PVMo/HMont are presented in Fig.3.The XRD pattern of PVMo shows its principal diffraction angles at 8.1, 8.9, 9.3, 27.8, 28.2, and 28.88° 2θ.This agrees with those reported in the literature59and correspond to 13-14 H 2 O hydrates, which are stable at room temperature.The PVMo/HMont materials exhibited refraction, which is due to a basal spacing of 1.80 nm. a

Table 3
Activity of various catalysts towards the oxidation of cyclohexene

Table 5
E a determined with different catalysts and different oxidants.

Table 6
Vanadium leaching determined by ICP analysis